Skip to contentSkip to conversation

schifra: fix build of examples and add hash

PATCH New Arnout Vandecappelle (Essensium/Mind) 3 participants 6 messages Oct 28, 2014 – Oct 31, 2014

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 schifra: fix build of examples and add hash

The OPTIONS argument passed to make had a redundant -o option, leading to:

g++ -pipe -Os -o -o schifra_reed_solomon_speed_evaluation schifra_reed_solomon_speed_evaluation.cpp -lstdc++ -lm
arm-linux-musleabi-g++: error: schifra_reed_solomon_speed_evaluation: No such file or directory

Fixes:
http://autobuild.buildroot.net/results/2143b4f6f22f50ccb4af36732cc7d9304daa3ff9/

This error didn't occur before because upstream has silently updated the tarball. To protect against this in the future, a hash is added as well.

---
 package/schifra/schifra.hash | 3 +++
 package/schifra/schifra.mk   | 2 +-
 2 files changed, 4 insertions(+), 1 deletion(-)
 create mode 100644 package/schifra/schifra.hash
package/schifra/schifra.hash new file +3 -0
new file mode 100644
index 0000000..7f8e6c0
@@ -0,0 +1,3 @@
+# Calculated locally after download
+# Upstream doesn't store a version and tarball may change unexpectedly!
+sha256	3e60478cf024d87735d1cc1460aa2176c449e6c7f76914eefda25d3b20d22522	schifra.tgz
package/schifra/schifra.mk +1 -1
index 96387d3..e55bf02 100644
@@ -12,7 +12,7 @@ SCHIFRA_LICENSE = schifra license
 SCHIFRA_LICENSE_FILES = schifra_license.txt
 
 SCHIFRA_MAKE_OPTS = COMPILER="$(TARGET_CXX)" \
-		   OPTIONS="$(TARGET_CFLAGS) $(TARGET_LDFLAGS) -o"
+		   OPTIONS="$(TARGET_CFLAGS) $(TARGET_LDFLAGS)"
 
 # The examples are the only buildable artefacts.
 ifeq ($(BR2_PACKAGE_SCHIFRA_EXAMPLES),y)
Signature · 1 line
2.1.1
#1 lore raw

PATCH v2 1/1 schifra: fix build of examples and add hash

Dear Arnout Vandecappelle (Essensium/Mind),

Quoted 13 lines · Arnout Vandecappelle (Essensium/Mind)

The OPTIONS argument passed to make had a redundant -o option, leading to:

g++ -pipe -Os -o -o schifra_reed_solomon_speed_evaluation schifra_reed_solomon_speed_evaluation.cpp -lstdc++ -lm
arm-linux-musleabi-g++: error: schifra_reed_solomon_speed_evaluation: No such file or directory

Fixes:
http://autobuild.buildroot.net/results/2143b4f6f22f50ccb4af36732cc7d9304daa3ff9/

This error didn't occur before because upstream has silently updated the tarball. To protect against this in the future, a hash is added as well.

Quoted 6 lines
---
 package/schifra/schifra.hash | 3 +++
 package/schifra/schifra.mk   | 2 +-
 2 files changed, 4 insertions(+), 1 deletion(-)
 create mode 100644 package/schifra/schifra.hash

Applied, thanks.

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
#2 lore raw

PATCH v2 1/1 schifra: fix build of examples and add hash

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Thomas" == Thomas Petazzoni <[email protected]> writes:

Hi,

Quoted 13 lines
Quoted 13 lines

The OPTIONS argument passed to make had a redundant -o option, leading to:

g++ -pipe -Os -o -o schifra_reed_solomon_speed_evaluation schifra_reed_solomon_speed_evaluation.cpp -lstdc++ -lm
arm-linux-musleabi-g++: error: schifra_reed_solomon_speed_evaluation: No such file or directory

Fixes:
http://autobuild.buildroot.net/results/2143b4f6f22f50ccb4af36732cc7d9304daa3ff9/

This error didn't occur before because upstream has silently updated the tarball. To protect against this in the future, a hash is added as well.

Quoted 8 lines
Quoted 6 lines

---
package/schifra/schifra.hash | 3 +++
package/schifra/schifra.mk | 2 +-

2 files changed, 4 insertions(+), 1 deletion(-)

create mode 100644 package/schifra/schifra.hash

Applied, thanks.

It seems like it changed AGAIN:

http://autobuild.buildroot.net/results/b62/b624c7f9fd7b431e97599e4b9b5e2a87fcb1a0a9/build-end.log

This is not workable. Nobody will be able to seriously use schifra in a project like this, and E.G. 2014.08 and earlier releases are broken.

What do we do?

- Get upstream to add a version number to the tarball and bump the version whenever it changes

- Put a known good version on sources.buildroot.net and use that

- Drop the package

Signature · 1 line
Bye, Peter Korsgaard
#3 lore raw

PATCH v2 1/1 schifra: fix build of examples and add hash

Dear Peter Korsgaard,

Quoted 6 lines · Peter Korsgaard

It seems like it changed AGAIN:

http://autobuild.buildroot.net/results/b62/b624c7f9fd7b431e97599e4b9b5e2a87fcb1a0a9/build-end.log

This is not workable. Nobody will be able to seriously use schifra in a project like this, and E.G. 2014.08 and earlier releases are broken.

Are we sure it changed again? Isn't the build error above caused by the fact that the tarball in the cache was wrong, and now that we've added the hash, the individual autobuilder instances are progressively realizing that the tarball they have in their cache is wrong.

Quoted 8 lines

What do we do?

- Get upstream to add a version number to the tarball and bump the version whenever it changes

- Put a known good version on sources.buildroot.net and use that

- Drop the package

I personally don't care about schifra. It was added back in August 2012 by Simon Dawson, but unfortunately Simon is no longer contributing to Buildroot.

Maybe we should first report the issue upstream?

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
#4 lore raw

PATCH v2 1/1 schifra: fix build of examples and add hash

Quoted 9 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Thomas" == Thomas Petazzoni <[email protected]> writes:

Dear Peter Korsgaard,

Quoted 4 lines · Peter Korsgaard

It seems like it changed AGAIN:

http://autobuild.buildroot.net/results/b62/b624c7f9fd7b431e97599e4b9b5e2a87fcb1a0a9/build-end.log

Quoted 7 lines
Quoted 2 lines

This is not workable. Nobody will be able to seriously use schifra in a project like this, and E.G. 2014.08 and earlier releases are broken.

Are we sure it changed again? Isn't the build error above caused by the fact that the tarball in the cache was wrong, and now that we've added the hash, the individual autobuilder instances are progressively realizing that the tarball they have in their cache is wrong.

Yes, I just tried downloading it here:
wget -q http://www.schifra.com/downloads/schifra.tgz
md5sum schifra.tgz 3b0aa73ed1ae1e2491e1a860a4f49f1b schifra.tgz

Which doesn't match the .hash file.

Quoted 10 lines
Quoted 6 lines

- Get upstream to add a version number to the tarball and bump the version whenever it changes

- Put a known good version on sources.buildroot.net and use that

- Drop the package

I personally don't care about schifra. It was added back in August 2012 by Simon Dawson, but unfortunately Simon is no longer contributing to Buildroot.

Indeed, I'm of the same opinion.

Quoted 1 line

Maybe we should first report the issue upstream?

I'll send a mail, and if it doesn't get resolved soon (or if somebody steps up to say they need it), then I'll simply drop the package.

Signature · 1 line
Bye, Peter Korsgaard
#5 lore raw

PATCH v2 1/1 schifra: fix build of examples and add hash

Dear Peter Korsgaard,

Quoted 4 lines · Peter Korsgaard
Quoted 1 line

Maybe we should first report the issue upstream?

I'll send a mail, and if it doesn't get resolved soon (or if somebody steps up to say they need it), then I'll simply drop the package.

Fine with me.

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com