Skip to contentSkip to conversation

Fix make check-docs on Windows

PATCH Superseded Johannes Schindelin via GitGitGadget 5 participants 17 messages Mar 13, 2019 – Mar 29, 2019

Subsystems (2)
Documentation Kbuild
lore
Share & tools
lore

Johannes Schindelin via GitGitGadget

author
#0 lore raw

cover Fix make check-docs on Windows

I stumbled across this when investigating one of Duy's bugs in the CI builds.

Johannes Schindelin (1):
check-docs: fix for setups where executables have an extension

 Makefile | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

base-commit: e902e9bcae2010bc42648c80ab6adc6c5a16a4a5
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-162%2Fdscho%2Fcheck-docs-on-windows-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-162/dscho/check-docs-on-windows-v1
Pull-Request: https://github.com/gitgitgadget/git/pull/162

Signature · 1 line
gitgitgadget
author
#1 lore raw

PATCH 1/1 check-docs: fix for setups where executables have an extension

From: Johannes Schindelin <[email protected]>

On Windows, for example, executables (must) have the extension .exe. Our check-docs target was not prepared for that.

Signed-off-by Johannes Schindelin <[email protected]>
---
 Makefile | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Makefile +2 -2
index 537493822b..df56bf0cd1 100644
@@ -3074,7 +3074,7 @@ ALL_COMMANDS += git-gui git-citool
 .PHONY: check-docs
 check-docs::
 	$(MAKE) -C Documentation lint-docs
-	@(for v in $(ALL_COMMANDS); \
+	@(for v in $(patsubst %$X,%,$(ALL_COMMANDS)); \
 	do \
 		case "$$v" in \
 		git-merge-octopus | git-merge-ours | git-merge-recursive | \

@@ -3103,7 +3103,7 @@ check-docs::
 		    -e 's/\.txt//'; \
 	) | while read how cmd; \
 	do \
-		case " $(ALL_COMMANDS) " in \
+		case "  $(patsubst %$X,%,$(ALL_COMMANDS)) " in \
 		*" $$cmd "*)	;; \
 		*) echo "removed but $$how: $$cmd" ;; \
 		esac; \
Signature · 1 line
gitgitgadget
#2 lore raw

Re: [PATCH 1/1] check-docs: fix for setups where executables have an extension

"Johannes Schindelin via GitGitGadget" <[email protected]>

Quoted 7 lines · writes:

From: Johannes Schindelin <[email protected]>

On Windows, for example, executables (must) have the extension .exe. Our check-docs target was not prepared for that.

Signed-off-by Johannes Schindelin <[email protected]>

---

My writing this loop certainly predates increased upstream activities for Windows port we see recently, and we saw little updates ever since the loop was written; I am kind of surprised it took this long for us to find the missing .exe support here.

Thanks.

Quoted 25 lines
 Makefile | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Makefile +2 -2
index 537493822b..df56bf0cd1 100644
@@ -3074,7 +3074,7 @@ ALL_COMMANDS += git-gui git-citool
 .PHONY: check-docs
 check-docs::
 	$(MAKE) -C Documentation lint-docs
-	@(for v in $(ALL_COMMANDS); \
+	@(for v in $(patsubst %$X,%,$(ALL_COMMANDS)); \
 	do \
 		case "$$v" in \
 		git-merge-octopus | git-merge-ours | git-merge-recursive | \

@@ -3103,7 +3103,7 @@ check-docs::
 		    -e 's/\.txt//'; \
 	) | while read how cmd; \
 	do \
-		case " $(ALL_COMMANDS) " in \
+		case "  $(patsubst %$X,%,$(ALL_COMMANDS)) " in \
 		*" $$cmd "*)	;; \
 		*) echo "removed but $$how: $$cmd" ;; \
 		esac; \
#3 lore raw

Re: [PATCH 1/1] check-docs: fix for setups where executables have an extension

In reply to Johannes Schindelin via GitGitGadget

Hi Johannes,

Quoted 24 lines · Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

On Windows, for example, executables (must) have the extension .exe. Our check-docs target was not prepared for that.

Signed-off-by Johannes Schindelin <[email protected]>
---
 Makefile | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Makefile +1 -1
index 537493822b..df56bf0cd1 100644
@@ -3074,7 +3074,7 @@ ALL_COMMANDS += git-gui git-citool
 .PHONY: check-docs
 check-docs::
 	$(MAKE) -C Documentation lint-docs
-	@(for v in $(ALL_COMMANDS); \
+	@(for v in $(patsubst %$X,%,$(ALL_COMMANDS)); \
 	do \
 		case "$$v" in \
 		git-merge-octopus | git-merge-ours | git-merge-recursive | \

Makefile check-docs:: · line 3106

Quoted 5 lines
Makefile +1 -1
@@ -3103,7 +3103,7 @@ check-docs::
 		    -e 's/\.txt//'; \
 	) | while read how cmd; \
 	do \
-		case " $(ALL_COMMANDS) " in \
+		case "  $(patsubst %$X,%,$(ALL_COMMANDS)) " in \

I'm a little late to reading this thread, but I was curious why there are now two spaces around the patsubst as opposed to the one around ALL_COMMANDS?

Quoted 5 lines
Makefile +0 -1
 		*" $$cmd "*)	;; \
 		*) echo "removed but $$how: $$cmd" ;; \
 		esac; \
--

gitgitgadget

Thanks,
Taylor

#4 lore raw

Re: [PATCH 1/1] check-docs: fix for setups where executables have an extension

Hi Taylor,

Quoted 26 lines · Taylor Blau
Quoted 25 lines · Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

On Windows, for example, executables (must) have the extension .exe. Our check-docs target was not prepared for that.

Signed-off-by Johannes Schindelin <[email protected]>
---
 Makefile | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Makefile +1 -1
index 537493822b..df56bf0cd1 100644
@@ -3074,7 +3074,7 @@ ALL_COMMANDS += git-gui git-citool
 .PHONY: check-docs
 check-docs::
 	$(MAKE) -C Documentation lint-docs
-	@(for v in $(ALL_COMMANDS); \
+	@(for v in $(patsubst %$X,%,$(ALL_COMMANDS)); \
 	do \
 		case "$$v" in \
 		git-merge-octopus | git-merge-ours | git-merge-recursive | \

@@ -3103,7 +3103,7 @@ check-docs::
 		    -e 's/\.txt//'; \
Quoted 8 lines

Makefile check-docs:: · line 3106

Quoted 4 lines
Makefile +1 -1
 	) | while read how cmd; \
 	do \
-		case " $(ALL_COMMANDS) " in \
+		case "  $(patsubst %$X,%,$(ALL_COMMANDS)) " in \

I'm a little late to reading this thread, but I was curious why there are now two spaces around the patsubst as opposed to the one around ALL_COMMANDS?

Sharp eyes, and a *very* good point. The double space is actually required for this patch to work as intended. I added the following explanation to the commit message:

Note that $(ALL_COMMANDS) starts with a space, and that is rather crucial for the while read how cmd loop: some of the input lines do not actually have the form <how> <cmd>, but only <cmd>, therefore $cmd evaluates to the empty string, and when we are looking for * $cmd * in $(ALL_COMMANDS), we do find it because the latter starts with a double space.

In other words, the double space helps us skip the lines that list only a command.

But in this patch, we want to transform $(ALL_COMMANDS) via $(patsubst ...) which swallows that double space. To help that, we prefix the result with *two* spaces instead of just one.

Do you think that is good enough?

Oh, and can I ask you to review *all* of my patches in the future?

Thanks,
Dscho

#5 lore raw

Re: [PATCH 1/1] check-docs: fix for setups where executables have an extension

Quoted 4 lines · Johannes Schindelin

Sharp eyes, and a *very* good point. The double space is actually required for this patch to work as intended. I added the following explanation to the commit message:

Quoted 8 lines

Note that $(ALL_COMMANDS) starts with a space, and that is rather crucial for the while read how cmd loop: some of the input lines do not actually have the form <how> <cmd>, but only <cmd>, therefore $cmd evaluates to the empty string, and when we are looking for * $cmd * in $(ALL_COMMANDS), we do find it because the latter starts with a double space.
In other words, the double space helps us skip the lines that list only a command.

That sounds more like a bug in the existing set-up even before your patch (in fact, these lines are probably from 2007 or so, long before you started touching our top-level Makefile heavily). If we want to ignore lines with only one token, I'd rather see it explicitly done, e.g.

	... generate data ... |
	while read how cmd
	do
		# skip a line with a single token
		case "$cmd" in "") continue ;; esac

		# is $cmd part of the known command list?
		case " $(ALL_COMMANDS) " in
		*" $cmd "*)	;; # skip
		*)	echo ... warning ... ;;
		esac
	done

instead of relying on "if $cmd is empty, then SP + $cmd + SP makes double space, so let's have double space somewhere in the list of known commands" cuteness.

Thanks.

#6 lore raw

Re: [PATCH 1/1] check-docs: fix for setups where executables have an extension

Hi Junio,

Quoted 34 lines · Junio C Hamano
Quoted 4 lines · Johannes Schindelin

Sharp eyes, and a *very* good point. The double space is actually required for this patch to work as intended. I added the following explanation to the commit message:

Quoted 8 lines

Note that $(ALL_COMMANDS) starts with a space, and that is rather crucial for the while read how cmd loop: some of the input lines do not actually have the form <how> <cmd>, but only <cmd>, therefore $cmd evaluates to the empty string, and when we are looking for * $cmd * in $(ALL_COMMANDS), we do find it because the latter starts with a double space.
In other words, the double space helps us skip the lines that list only a command.

That sounds more like a bug in the existing set-up even before your patch (in fact, these lines are probably from 2007 or so, long before you started touching our top-level Makefile heavily). If we want to ignore lines with only one token, I'd rather see it explicitly done, e.g.

	... generate data ... |
	while read how cmd
	do
		# skip a line with a single token
		case "$cmd" in "") continue ;; esac

		# is $cmd part of the known command list?
		case " $(ALL_COMMANDS) " in
		*" $cmd "*)	;; # skip
		*)	echo ... warning ... ;;
		esac
	done
Quoted 3 lines

instead of relying on "if $cmd is empty, then SP + $cmd + SP makes double space, so let's have double space somewhere in the list of known commands" cuteness.

You are right, I should have root-caused it, it had a funny smell to it.

Turns out that we should not just skip those lines with a single token, but instead fix the bug that prevents the how variable to be set to documented in those cases, as had been intended all along.

I fixed the commit, and accompanied it with a bug fix for this, plus for the fall-out bugs that had been hidden by this bug.

Ciao,
Dscho

author
#7 lore raw

v2 cover Fix make check-docs on Windows

In reply to Johannes Schindelin via GitGitGadget

I stumbled across this when investigating one of Duy's bugs in the CI builds.

Changes since v1:

* We no longer add an extra space in front of the $(patsubst ...) in the second loop, but fix the underlying bug that prevented check-docs' part from working where it discovered documented commands that are actually not installed.
* As a fall-out, the gitremote-helpers page, which does not refer to a command, but to a concept, was moved from section 1 to section 7.
* As a second fall-out, the documentation for git remote-testgit, which is a command, but not installed, was removed.
* As a "while at it" patch, the part of the check-docs target that tried to identify commands that are listed in the help, but not implemented, was fixed to no longer mistake guides like gitattributes and gitcli for commands.

Johannes Schindelin (5):
docs: move gitremote-helpers into section 7
docs: do not document the git remote-testgit command
check-docs: really look at the documented commands again
check-docs: do not expect guide pages to correspond to commands
check-docs: fix for setups where executables have an extension

 Documentation/Makefile                |  2 +-
 Documentation/git-remote-ext.txt      |  2 +-
 Documentation/git-remote-fd.txt       |  2 +-
 Documentation/git-remote-helpers.txto |  2 +-
 Documentation/git-remote-testgit.txt  | 30 ---------------------------
 Documentation/gitremote-helpers.txt   |  4 +---
 Documentation/urls.txt                |  2 +-
 Makefile                              |  7 ++++---
 8 files changed, 10 insertions(+), 41 deletions(-)
 delete mode 100644 Documentation/git-remote-testgit.txt

base-commit: e902e9bcae2010bc42648c80ab6adc6c5a16a4a5
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-162%2Fdscho%2Fcheck-docs-on-windows-v2
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-162/dscho/check-docs-on-windows-v2
Pull-Request: https://github.com/gitgitgadget/git/pull/162

Range-diff vs v1:

 -:  ---------- > 1:  0ad38649c0 docs: move gitremote-helpers into section 7
 -:  ---------- > 2:  810d2c5a94 docs: do not document the `git remote-testgit` command
 -:  ---------- > 3:  0dad6abd2f check-docs: really look at the documented commands again
 -:  ---------- > 4:  1097be7678 check-docs: do not expect guide pages to correspond to commands
 1:  f06126c3a1 ! 5:  b26aa40c61 check-docs: fix for setups where executables have an extension
     @@ -24,7 +24,7 @@
       	) | while read how cmd; \
       	do \
      -		case " $(ALL_COMMANDS) " in \
     -+		case "  $(patsubst %$X,%,$(ALL_COMMANDS)) " in \
     ++		case " $(patsubst %$X,%,$(ALL_COMMANDS)) " in \
       		*" $$cmd "*)	;; \
       		*) echo "removed but $$how: $$cmd" ;; \
       		esac; \
Signature · 1 line
gitgitgadget
author
#13 lore raw

PATCH v2 3/5 check-docs: really look at the documented commands again

From: Johannes Schindelin <[email protected]>

As part of the check-docs target, we verify that commands that are documented are actually in the current list of commands to be built.

However, this logic broke in 5fafce0b78 (check-docs: get documented command list from Makefile, 2012-08-08), when we tried to make the logic safer by not looking at the files in the worktree, but at the list of files to be generated in Documentation/Makefile. While this was the right thing to do, it failed to accommodate for the fact that make -C Documentation/ print-man1, unlike ls Documentation/*.txt, would *not* print lines starting with the prefix Documentation/.

At long last, let's fix this.

Note: This went undetected due to a funny side effect of the ALL_PROGRAMS variable starting with a space. That space, together with the extra space we inserted before $(ALL_PROGRAMS) in the

	case " $(ALL_PROGRAMS)" in
	*" $$cmd ")
		[...]

construct, is responsible that this case arm is used when cmd is empty (which was clearly not intended to be the case).

Signed-off-by Johannes Schindelin <[email protected]>
---
 Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Makefile +1 -1
index 537493822b..76904f07fe 100644
@@ -3099,7 +3099,7 @@ check-docs::
 		    -e 's/^/listed /' command-list.txt; \
 		$(MAKE) -C Documentation print-man1 | \
 		grep '\.txt$$' | \
-		sed -e 's|Documentation/|documented |' \
+		sed -e 's|^|documented |' \
 		    -e 's/\.txt//'; \
 	) | while read how cmd; \
 	do \
Signature · 1 line
gitgitgadget
author
#15 lore raw

PATCH v2 5/5 check-docs: fix for setups where executables have an extension

In reply to Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

On Windows, for example, executables (must) have the extension .exe. Our check-docs target was not prepared for that.

Signed-off-by Johannes Schindelin <[email protected]>
---
 Makefile | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Makefile +2 -2
index c7b884cd38..2cd650e7b9 100644
@@ -3074,7 +3074,7 @@ ALL_COMMANDS += git-gui git-citool
 .PHONY: check-docs
 check-docs::
 	$(MAKE) -C Documentation lint-docs
-	@(for v in $(ALL_COMMANDS); \
+	@(for v in $(patsubst %$X,%,$(ALL_COMMANDS)); \
 	do \
 		case "$$v" in \
 		git-merge-octopus | git-merge-ours | git-merge-recursive | \

@@ -3104,7 +3104,7 @@ check-docs::
 		    -e 's/\.txt//'; \
 	) | while read how cmd; \
 	do \
-		case " $(ALL_COMMANDS) " in \
+		case " $(patsubst %$X,%,$(ALL_COMMANDS)) " in \
 		*" $$cmd "*)	;; \
 		*) echo "removed but $$how: $$cmd" ;; \
 		esac; \
Signature · 1 line
gitgitgadget
author
#9 lore raw

PATCH v2 2/5 docs: do not document the `git remote-testgit` command

In reply to Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

Since 7ded055401 (build: do not install git-remote-testgit, 2013-06-07), we do not install it. Therefore it makes no sense to document it, either.

Signed-off-by Johannes Schindelin <[email protected]>
---
 Documentation/git-remote-testgit.txt | 30 ----------------------------
 Documentation/gitremote-helpers.txt  |  2 --
 2 files changed, 32 deletions(-)
 delete mode 100644 Documentation/git-remote-testgit.txt
Documentation/git-remote-testgit.txt deleted +0 -30
deleted file mode 100644
index b45bfebba5..0000000000
@@ -1,30 +0,0 @@
-git-remote-testgit(1)
-=====================
-
-NAME
-----
-git-remote-testgit - Example remote-helper
-
-
-SYNOPSIS
---------
-[verse]
-git clone testgit::<source-repo> [<destination>]
-
-DESCRIPTION
------------
-
-This command is a simple remote-helper, that is used both as a
-testcase for the remote-helper functionality, and as an example to
-show remote-helper authors one possible implementation.
-
-The best way to learn more is to read the comments and source code in
-'git-remote-testgit'.
-
-SEE ALSO
---------
-linkgit:gitremote-helpers[7]
-
-GIT
----
-Part of the linkgit:git[1] suite
Documentation/gitremote-helpers.txt +0 -2
index 34a3e60d08..2fc4007525 100644
@@ -513,8 +513,6 @@ linkgit:git-remote-ext[1]
 
 linkgit:git-remote-fd[1]
 
-linkgit:git-remote-testgit[1]
-
 linkgit:git-fast-import[1]
 
 GIT
Signature · 1 line
gitgitgadget
author
#8 lore raw

PATCH v2 1/5 docs: move gitremote-helpers into section 7

In reply to Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

It is currently in section 1, but that section is intended for "Executable programs or shell commands".

A more appropriate place is section 7: "Miscellaneous (including macro packages and conventions), e.g. man(7), groff(7)".

This issue should have been detected earlier by make check-docs, but was missed due to a bug in that Makefile target (that we are about to fix).

Signed-off-by Johannes Schindelin <[email protected]>
---
 Documentation/Makefile                | 2 +-
 Documentation/git-remote-ext.txt      | 2 +-
 Documentation/git-remote-fd.txt       | 2 +-
 Documentation/git-remote-helpers.txto | 2 +-
 Documentation/git-remote-testgit.txt  | 2 +-
 Documentation/gitremote-helpers.txt   | 2 +-
 Documentation/urls.txt                | 2 +-
 7 files changed, 7 insertions(+), 7 deletions(-)
Documentation/Makefile +1 -1
index 26a2342bea..af0e2cf11a 100644
@@ -12,7 +12,6 @@ MAN1_TXT += $(filter-out \
 		$(wildcard git-*.txt))
 MAN1_TXT += git.txt
 MAN1_TXT += gitk.txt
-MAN1_TXT += gitremote-helpers.txt
 MAN1_TXT += gitweb.txt
 
 MAN5_TXT += gitattributes.txt

@@ -30,6 +29,7 @@ MAN7_TXT += gitdiffcore.txt
 MAN7_TXT += giteveryday.txt
 MAN7_TXT += gitglossary.txt
 MAN7_TXT += gitnamespaces.txt
+MAN7_TXT += gitremote-helpers.txt
 MAN7_TXT += gitrevisions.txt
 MAN7_TXT += gitsubmodules.txt
 MAN7_TXT += gittutorial-2.txt
Documentation/git-remote-ext.txt +1 -1
index 3fc5d94336..790e2ca3fe 100644
@@ -118,7 +118,7 @@ begins with `ext::`.  Examples:
 
 SEE ALSO
 --------
-linkgit:gitremote-helpers[1]
+linkgit:gitremote-helpers[7]
 
 GIT
 ---
Documentation/git-remote-fd.txt +1 -1
index 80afca866c..0451ceb8a2 100644
@@ -52,7 +52,7 @@ EXAMPLES
 
 SEE ALSO
 --------
-linkgit:gitremote-helpers[1]
+linkgit:gitremote-helpers[7]
 
 GIT
 ---
Documentation/git-remote-helpers.txto +1 -1
index 49233f5d26..6f353ebfd3 100644
@@ -1,7 +1,7 @@
 git-remote-helpers
 ==================
 
-This document has been moved to linkgit:gitremote-helpers[1].
+This document has been moved to linkgit:gitremote-helpers[7].
 
 Please let the owners of the referring site know so that they can update the
 link you clicked to get here.
Documentation/git-remote-testgit.txt +1 -1
index f791d73c05..b45bfebba5 100644
@@ -23,7 +23,7 @@ The best way to learn more is to read the comments and source code in
 
 SEE ALSO
 --------
-linkgit:gitremote-helpers[1]
+linkgit:gitremote-helpers[7]
 
 GIT
 ---
Documentation/gitremote-helpers.txt +1 -1
index 9d1459aac6..34a3e60d08 100644
@@ -1,4 +1,4 @@
-gitremote-helpers(1)
+gitremote-helpers(7)
 ====================
 
 NAME
Documentation/urls.txt +1 -1
index b05da95788..bc354fe2dc 100644
@@ -62,7 +62,7 @@ may be used:
 
 where <address> may be a path, a server and path, or an arbitrary
 URL-like string recognized by the specific remote helper being
-invoked. See linkgit:gitremote-helpers[1] for details.
+invoked. See linkgit:gitremote-helpers[7] for details.
 
 If there are a large number of similarly-named remote repositories and
 you want to use a different format for them (such that the URLs you
Signature · 1 line
gitgitgadget
author
#14 lore raw

PATCH v2 4/5 check-docs: do not expect guide pages to correspond to commands

In reply to Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

When we want to see what commands are listed in command-list.txt but not installed, we currently include lines that refer to guides, e.g. gitattributes or gitcli.

Let's not include those lines, as they are not referring to commands.

Signed-off-by Johannes Schindelin <[email protected]>
---
 Makefile | 1 +
 1 file changed, 1 insertion(+)
Makefile +1 -0
index 76904f07fe..c7b884cd38 100644
@@ -3095,6 +3095,7 @@ check-docs::
 	( \
 		sed -e '1,/^### command list/d' \
 		    -e '/^#/d' \
+		    -e '/guide$$/d' \
 		    -e 's/[ 	].*//' \
 		    -e 's/^/listed /' command-list.txt; \
 		$(MAKE) -C Documentation print-man1 | \
Signature · 1 line
gitgitgadget
#16 lore raw
In reply to Johannes Schindelin via GitGitGadget

Hi,

Quoted 6 lines · Johannes Schindelin via GitGitGadget

Johannes Schindelin (5):
docs: move gitremote-helpers into section 7
docs: do not document the git remote-testgit command
check-docs: really look at the documented commands again
check-docs: do not expect guide pages to correspond to commands
check-docs: fix for setups where executables have an extension

FWIW I am also working on a follow-up patch series [*1*] that resolves all issues reported by make check-docs on Windows, but I do not want to make this here patch series even bigger (from an initial single, relatively small patch).

Ciao,
Johannes

Foontote *1*: Tracked here: https://github.com/gitgitgadget/git/pull/168

#10 lore raw

Re: [PATCH v2 2/5] docs: do not document the `git remote-testgit` command

In reply to Johannes Schindelin via GitGitGadget
Quoted 54 lines · Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

Since 7ded055401 (build: do not install git-remote-testgit, 2013-06-07), we do not install it. Therefore it makes no sense to document it, either.

Signed-off-by Johannes Schindelin <[email protected]>
---
 Documentation/git-remote-testgit.txt | 30 ----------------------------
 Documentation/gitremote-helpers.txt  |  2 --
 2 files changed, 32 deletions(-)
 delete mode 100644 Documentation/git-remote-testgit.txt
Documentation/git-remote-testgit.txt deleted +0 -30
deleted file mode 100644
index b45bfebba5..0000000000
@@ -1,30 +0,0 @@
-git-remote-testgit(1)
-=====================
-
-NAME
-----
-git-remote-testgit - Example remote-helper
-
-
-SYNOPSIS
---------
-[verse]
-git clone testgit::<source-repo> [<destination>]
-
-DESCRIPTION
------------
-
-This command is a simple remote-helper, that is used both as a
-testcase for the remote-helper functionality, and as an example to
-show remote-helper authors one possible implementation.
-
-The best way to learn more is to read the comments and source code in
-'git-remote-testgit'.
-
-SEE ALSO
---------
-linkgit:gitremote-helpers[7]
-
-GIT
----
-Part of the linkgit:git[1] suite
Documentation/gitremote-helpers.txt +0 -0
index 34a3e60d08..2fc4007525 100644

Documentation/gitremote-helpers.txt linkgit:git-remote-ext[1] · line 516

Quoted 6 lines
Documentation/gitremote-helpers.txt +0 -2
@@ -513,8 +513,6 @@ linkgit:git-remote-ext[1]
 
 linkgit:git-remote-fd[1]
 
-linkgit:git-remote-testgit[1]
-
 linkgit:git-fast-import[1]

I wonder if it should be moved into e.g. t/helper or somesuch at this point...

#11 lore raw

Re: [PATCH v2 2/5] docs: do not document the `git remote-testgit` command

Hi Ævar,

Quoted 58 lines · Ævar Arnfjörð Bjarmason
Quoted 55 lines · Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

Since 7ded055401 (build: do not install git-remote-testgit, 2013-06-07), we do not install it. Therefore it makes no sense to document it, either.

Signed-off-by Johannes Schindelin <[email protected]>
---
 Documentation/git-remote-testgit.txt | 30 ----------------------------
 Documentation/gitremote-helpers.txt  |  2 --
 2 files changed, 32 deletions(-)
 delete mode 100644 Documentation/git-remote-testgit.txt
Documentation/git-remote-testgit.txt deleted +0 -30
deleted file mode 100644
index b45bfebba5..0000000000
@@ -1,30 +0,0 @@
-git-remote-testgit(1)
-=====================
-
-NAME
-----
-git-remote-testgit - Example remote-helper
-
-
-SYNOPSIS
---------
-[verse]
-git clone testgit::<source-repo> [<destination>]
-
-DESCRIPTION
------------
-
-This command is a simple remote-helper, that is used both as a
-testcase for the remote-helper functionality, and as an example to
-show remote-helper authors one possible implementation.
-
-The best way to learn more is to read the comments and source code in
-'git-remote-testgit'.
-
-SEE ALSO
---------
-linkgit:gitremote-helpers[7]
-
-GIT
----
-Part of the linkgit:git[1] suite
Documentation/gitremote-helpers.txt +0 -0
index 34a3e60d08..2fc4007525 100644
@@ -513,8 +513,6 @@ linkgit:git-remote-ext[1]
Quoted 8 lines

Documentation/gitremote-helpers.txt linkgit:git-remote-ext[1] · line 516

Quoted 5 lines
Documentation/gitremote-helpers.txt +0 -2
 linkgit:git-remote-fd[1]
 
-linkgit:git-remote-testgit[1]
-
 linkgit:git-fast-import[1]

I wonder if it should be moved into e.g. t/helper or somesuch at this point...

Sure.

One thing to keep in mind is that we cannot change the actual name of the executable, as it needs to have the format git-remote-*, and cannot require a first argument testgit, to be able to perform its purpose.

We do have a precedent for that kind of scenario: test-fake-ssh. It is also not folded into test-tool, and it is copied while changing the name to ssh, as required to fulfill its mission.

I am not totally enthusiastic about this because this is what prevents Git's test suite from running in a RAM disk on Windows, IIRC: I have not been able to convince my RAM disk provider to allow executables stored on it to be executed.

Ciao,
Dscho

#12 lore raw

Re: [PATCH v2 2/5] docs: do not document the `git remote-testgit` command

Hi Ævar,

Quoted 74 lines · Johannes Schindelin
Quoted 57 lines · Ævar Arnfjörð Bjarmason
Quoted 55 lines · Johannes Schindelin via GitGitGadget

From: Johannes Schindelin <[email protected]>

Since 7ded055401 (build: do not install git-remote-testgit, 2013-06-07), we do not install it. Therefore it makes no sense to document it, either.

Signed-off-by Johannes Schindelin <[email protected]>
---
 Documentation/git-remote-testgit.txt | 30 ----------------------------
 Documentation/gitremote-helpers.txt  |  2 --
 2 files changed, 32 deletions(-)
 delete mode 100644 Documentation/git-remote-testgit.txt
Documentation/git-remote-testgit.txt deleted +0 -30
deleted file mode 100644
index b45bfebba5..0000000000
@@ -1,30 +0,0 @@
-git-remote-testgit(1)
-=====================
-
-NAME
-----
-git-remote-testgit - Example remote-helper
-
-
-SYNOPSIS
---------
-[verse]
-git clone testgit::<source-repo> [<destination>]
-
-DESCRIPTION
------------
-
-This command is a simple remote-helper, that is used both as a
-testcase for the remote-helper functionality, and as an example to
-show remote-helper authors one possible implementation.
-
-The best way to learn more is to read the comments and source code in
-'git-remote-testgit'.
-
-SEE ALSO
---------
-linkgit:gitremote-helpers[7]
-
-GIT
----
-Part of the linkgit:git[1] suite
Documentation/gitremote-helpers.txt +0 -0
index 34a3e60d08..2fc4007525 100644
@@ -513,8 +513,6 @@ linkgit:git-remote-ext[1]
Quoted 8 lines

Documentation/gitremote-helpers.txt linkgit:git-remote-ext[1] · line 516

Quoted 5 lines
Documentation/gitremote-helpers.txt +0 -2
 linkgit:git-remote-fd[1]
 
-linkgit:git-remote-testgit[1]
-
 linkgit:git-fast-import[1]

I wonder if it should be moved into e.g. t/helper or somesuch at this point...

Sure.

One thing to keep in mind is that we cannot change the actual name of the executable, as it needs to have the format git-remote-*, and cannot require a first argument testgit, to be able to perform its purpose.

Quoted 8 lines

We do have a precedent for that kind of scenario: test-fake-ssh. It is also not folded into test-tool, and it is copied while changing the name to ssh, as required to fulfill its mission.

I am not totally enthusiastic about this because this is what prevents Git's test suite from running in a RAM disk on Windows, IIRC: I have not been able to convince my RAM disk provider to allow executables stored on it to be executed.

Oh you know what? git-remote-testgit is a *shell script*.

So that one was easy...

Now, the question is really, should I just add this patch *here*, or to the non-Windows-related docs fixes I have lined up?

Decisions, decisions.

Ciao,
Dscho