Conversation
|
ping @GabrielBrascher @nathanejohnson , this slipped in end of march. didn't notice, too much on fedora :(, not enough on a mac :) |
| sed -e 's,%API_HEADER%,All APIs,g' "$thisdir/generatetoc_header.xsl" >generatetoc.xsl | ||
| sed -i "s/%ACS_RELEASE%/${ACS_RELEASE}/g" generatetoc.xsl | ||
| sed -i "s/%ACS_RELEASE%/${ACS_RELEASE}/g" generatecommands.xsl | ||
| sed -i .bak "s/%ACS_RELEASE%/${ACS_RELEASE}/g" generatetoc.xsl |
There was a problem hiding this comment.
What if we execute the commands at line 63 and 64 like the one in 62? Or, maybe the other way around.
I mean, they do the same thing, but with different approaches. Therefore, it might be interesting to maintain some consistency here.
There was a problem hiding this comment.
@rafaelweingartner, not entirely sure what you mean but 62 an 63 can certainly be combined. not sure how 64 can, it is touching a different file.
There was a problem hiding this comment.
I meant the following:
All of the commands are replacing "place holders" in the same file. However, two of them are doing in line (via -i flag), whereas line 62 is replacing the entire file (not using the -i flag and redirecting the output of the sed to a file with the same name). I do not mind if we do in different commands the replace, but we could at least try to use the same approach in all of them.
By approach here I mean the method, either inline or not.
There was a problem hiding this comment.
I think you are misreading the code,
line 62 redirects to a new file, from "$thisdir/generatetoc_header.xsl" to generatetoc.xsl
then 63 operates on the resulting generatetoc.xsl
but 64 is operating on generatecommands.xsl
I don't think shell operations would allow redecting to input, it would empty the file before openening it for read if you tried that.
There was a problem hiding this comment.
@DaanHoogland @rafaelweingartner A (ugly) solution is running the command with the -e and move the new file to the old one:
sed -e "s/%ACS_RELEASE%/${ACS_RELEASE}/g" newGeneratetoc.xsl && mv newGeneratetoc.xsl generatetoc.xsl;
There was a problem hiding this comment.
@DaanHoogland on linux the -i must be appended with the .bak (-i.bak)
-i[SUFFIX], --in-place[=SUFFIX]
edit files in place (makes backup if SUFFIX supplied)
|
@DaanHoogland env issue, the sed on OSX is freebsd based you can try gnu-sed like: |
I realize that, but it complicates build and thus inhibits people entering the playing field. I would prefer adding the extension, it is in the target dir and doesn't hurt anybody. |
|
@rhtyd you can easily add gnu sed to mac via homebrew as you point out, however the binary is called 'gsed' to avoid conflicting with the main sed installed in /usr/bin . |
|
@blueorangutan package |
|
@rhtyd a Jenkins job has been kicked to build packages. I'll keep you posted as I make progress. |
|
Packaging result: ✖centos6 ✖centos7 ✖debian. JID-2752 |
|
@DaanHoogland I don't think that this would work on linux. The Documentation: |
|
closed-open PR to rekick Travis |
|
@rhtyd a Jenkins job has been kicked to build packages. I'll keep you posted as I make progress. |
|
|
Packaging result: ✖centos6 ✖centos7 ✖debian. JID-2799 |
|
travis is quite clear in this case: needs tlc |
|
@DaanHoogland can you debug via the backend jenkins what failed? |
|
@rhtyd if you mean travis, yes. but kind of high prio -3. |
|
@DaanHoogland fails on linux with: |
|
This is |
|
Closing as #4253 fixed it |
Description
build fails on mac due to #3247 . adding extension that will only be created in '/target/' anyway
Types of changes
Screenshots (if appropriate):
How Has This Been Tested?