Michael J Gruber [off-list ref] writes:
When installing html-doc, install-webdoc.sh compares the installed html
with the version to be installed using diff. Currently, the diff output
fills up stdout.
As I've been using this output as the final sanity check on k.org
environment, I personally do not want to lose the output from that
particular diff.
When I push to the public repository, a hook is triggered to build and
install the documentation, sending its progress to a log file, and I have
a window open running "tail -f" on it.
Actually why don't we do the attached instead as the first step?
The timestamp that follows "Last updated " is formatted differently
depending on the version of AsciiDoc. Looking at 4604fe56 on "html"
branch, you can see that AsciiDoc 7.0.2 used to give "02-Jul-2008 03:02:14
UTC" but AsciiDoc 8.2.5 gave "2008-09-19 06:33:25 UTC". We haven't been
correctly filtering out phantom changes that result from only the build
date for some time now, it seems.
Side note:
How can you find the above easily? Here is one way, if you run git
with f506b8e (git log/diff: add -G<regexp> that greps in the patch
text, 2010-08-23) currently parked in 'pu':
git log -p -G"Last updated [0-9][0-9]-[A-Z][a-z][a-z]-" origin/html
Documentation/install-webdoc.sh | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/Documentation/install-webdoc.sh b/Documentation/install-webdoc.sh
index 34d02a2..37e67d1 100755
--- a/Documentation/install-webdoc.sh
+++ b/Documentation/install-webdoc.sh
@@ -12,7 +12,7 @@ do
then
: did not match
elif test -f "$T/$h" &&
- $DIFF -u -I'Last updated [0-9][0-9]-[A-Z][a-z][a-z]-' "$T/$h" "$h"
+ $DIFF -u -I'^Last updated ' "$T/$h" "$h"
then
:; # up to date
else
Junio C Hamano venit, vidit, dixit 01.09.2010 18:24:
Michael J Gruber [off-list ref] writes:
quoted
When installing html-doc, install-webdoc.sh compares the installed html
with the version to be installed using diff. Currently, the diff output
fills up stdout.
As I've been using this output as the final sanity check on k.org
environment, I personally do not want to lose the output from that
particular diff.
When I push to the public repository, a hook is triggered to build and
install the documentation, sending its progress to a log file, and I have
a window open running "tail -f" on it.
Sorry, isn't that a somewhat crazy type of sanity check? You're getting
the complete diff for the generated htmls flying by. Wouldn't the actual
"install..." be sufficient, or maybe, a QUIET_DIFF variant which
displays the command line? And not only sufficient, but actually easier
to check visually as you seem to do?
[A more throrough check would be comparing the name-diff of
Documentation/*.txt with that of the generated html.]
quoted hunk
Actually why don't we do the attached instead as the first step?
The timestamp that follows "Last updated " is formatted differently
depending on the version of AsciiDoc. Looking at 4604fe56 on "html"
branch, you can see that AsciiDoc 7.0.2 used to give "02-Jul-2008 03:02:14
UTC" but AsciiDoc 8.2.5 gave "2008-09-19 06:33:25 UTC". We haven't been
correctly filtering out phantom changes that result from only the build
date for some time now, it seems.
Side note:
How can you find the above easily? Here is one way, if you run git
with f506b8e (git log/diff: add -G<regexp> that greps in the patch
text, 2010-08-23) currently parked in 'pu':
git log -p -G"Last updated [0-9][0-9]-[A-Z][a-z][a-z]-" origin/html
Documentation/install-webdoc.sh | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/Documentation/install-webdoc.sh b/Documentation/install-webdoc.sh
index 34d02a2..37e67d1 100755
--- a/Documentation/install-webdoc.sh
+++ b/Documentation/install-webdoc.sh
@@ -12,7 +12,7 @@ do
then
: did not match
elif test -f "$T/$h" &&
- $DIFF -u -I'Last updated [0-9][0-9]-[A-Z][a-z][a-z]-' "$T/$h" "$h"
+ $DIFF -u -I'^Last updated ' "$T/$h" "$h"
then
:; # up to date
else
I think that is the right fix for the problem you mentioned, which is a
different one ;)
With your fix, the problem I'm after occurs less often, of course, maybe
to the extent of not being a real problem any more.
Michael