Broken link and defunct search

Broken link and defunct search

[Nikolay Mihaylov]

Hello

My sincere apologies if this is not the right place to report the 
problem but (and I quote from http://gcc.gnu.org/bugs.html):

Please send comments on these web pages and the development of GCC to 
our developer mailing list at gcc@gnu.org <mailto:gcc@gnu.org> or 
gcc@gcc.gnu.org <mailto:gcc@gcc.gnu.org>.

At the bottom of the page that I quote from there's a reference to 
"Common Misunderstandings with GNU C++" 
(http://gcc.gnu.org/onlinedocs/gcc/C---Misunderstandings.html) but 
there's no page at this URL. I tried to locate the page by searching for 
"Common Misunderstandings" at gnu.org but it finds no matches. It knows 
nothing about GNU or Richard Stallman (or anything else) either which is 
quite suspicious. Actually, the search at gcc.gnu.org works fine, albeit 
a bit slow, and gave me 
http://gcc.gnu.org/onlinedocs/gcc/C_002b_002b-Misunderstandings.html, 
which I guess is the correct link.

Kudos to all GNU developers for the great tools.

Nikolay

Re: Broken link and defunct search

[Janis Johnson]

On Sun, Aug 15, 2004 at 11:15:51AM +0200, Nikolay Mihaylov wrote:
> Hello
> 
> My sincere apologies if this is not the right place to report the 
> problem but (and I quote from http://gcc.gnu.org/bugs.html):
> 
> Please send comments on these web pages and the development of GCC to 
> our developer mailing list at gcc@gnu.org <mailto:gcc@gnu.org> or 
> gcc@gcc.gnu.org <mailto:gcc@gcc.gnu.org>.
> 
> At the bottom of the page that I quote from there's a reference to 
> "Common Misunderstandings with GNU C++" 
> (http://gcc.gnu.org/onlinedocs/gcc/C---Misunderstandings.html) but 
> there's no page at this URL. I tried to locate the page by searching for 
> "Common Misunderstandings" at gnu.org but it finds no matches. It knows 
> nothing about GNU or Richard Stallman (or anything else) either which is 
> quite suspicious. Actually, the search at gcc.gnu.org works fine, albeit 
> a bit slow, and gave me 
> http://gcc.gnu.org/onlinedocs/gcc/C_002b_002b-Misunderstandings.html, 
> which I guess is the correct link.

Nikolay, thanks for reporting this.

Gerald, the page in question is for node "C++ Misunderstandings".
makeinfo 4.7 replaces '+' in a file name with '_002b', and makeinfo
4.6 replaced it with '-'.  I don't know if the obvious fix of
changing the link in bugs.html is really the correct thing to do.

Janis

RE: Broken link and defunct search

[Dave Korn]

> -----Original Message-----
> From: gcc-owner On Behalf Of Janis Johnson
> Sent: 16 August 2004 21:32

> > At the bottom of the page that I quote from there's a reference to 
> > "Common Misunderstandings with GNU C++" 
> > (http://gcc.gnu.org/onlinedocs/gcc/C---Misunderstandings.html) but 
> > there's no page at this URL. I tried to locate the page by 

> http://gcc.gnu.org/onlinedocs/gcc/C_002b_002b-Misunderstandings.html, 
> > which I guess is the correct link.

> Gerald, the page in question is for node "C++ Misunderstandings".
> makeinfo 4.7 replaces '+' in a file name with '_002b', and makeinfo
> 4.6 replaced it with '-'.  I don't know if the obvious fix of
> changing the link in bugs.html is really the correct thing to do.
> 
> Janis

  We've just done this one!

http://gcc.gnu.org/ml/gcc/2004-08/msg00487.html
http://gcc.gnu.org/ml/gcc/2004-07/msg00946.html


  I guess there must be further instances scattered throughout the website.
It should be possible to dig them all up fairly easily by checking out the
repository with the webpages in and using "find . -type f | xargs grep
__00", shouldn't it?


    cheers, 
      DaveK
-- 
Can't think of a witty .sigline today....
 

RE: Broken link and defunct search

[Gerald Pfeifer]

On Tue, 17 Aug 2004, Dave Korn wrote:
> Janis wrote:
>> Gerald, the page in question is for node "C++ Misunderstandings".
>> makeinfo 4.7 replaces '+' in a file name with '_002b', and makeinfo
>> 4.6 replaced it with '-'.  I don't know if the obvious fix of
>> changing the link in bugs.html is really the correct thing to do.
> http://gcc.gnu.org/ml/gcc/2004-08/msg00487.html
> http://gcc.gnu.org/ml/gcc/2004-07/msg00946.html
> [...]
> I guess there must be further instances scattered throughout the website.
> It should be possible to dig them all up fairly easily by checking out the
> repository with the webpages in and using "find . -type f | xargs grep
> __00", shouldn't it?

Personally, I consider this an annoying bug in makeinfo 4.7; for user
visible addresses "_002b" is just ridiculous to use instead of "-", and
in our case possible conflicts between different strings simply are not
an issue.

I just sent a message to the texinfo maintainers to see whether they are 
willing to provide an option to disable this new behavior (or whether 
there is some workaround).

In the worst case, I'll probably do a big sed over the GCC web pages. :-(

Gerald

RE: Broken link and defunct search

[Gerald Pfeifer]

On Wed, 18 Aug 2004, Gerald Pfeifer wrote:
> Personally, I consider this an annoying bug in makeinfo 4.7; for user
> visible addresses "_002b" is just ridiculous to use instead of "-", and
> in our case possible conflicts between different strings simply are not
> an issue.
>
> I just sent a message to the texinfo maintainers to see whether they are 
> willing to provide an option to disable this new behavior (or whether there 
> is some workaround).

I have not received any response yet...

> In the worst case, I'll probably do a big sed over the GCC web pages. :-(

...so I just declared the worst case and added

   sed -e 's/_002d/-/g' -e 's/_002a/*/g'

to a strategic place to the wwwdocs/bin/preprocess script on gcc.gnu.org.

Yes, this is kind of embarrassing, but at least deep links into our
installation notes work again.


Joseph, do you want to do something similar for /onlinedocs?  The URLs
generated by makeinfo 4.7 are simply gross (and currently external links
and bookmarks into /onlinedocs are broken).

Gerald
-- 
Gerald Pfeifer (Jerry)   gerald@pfeifer.com   http://www.pfeifer.com/gerald/

RE: Broken link and defunct search

[Joseph S. Myers]

On Thu, 26 Aug 2004, Gerald Pfeifer wrote:

> ...so I just declared the worst case and added
> 
>   sed -e 's/_002d/-/g' -e 's/_002a/*/g'
> 
> to a strategic place to the wwwdocs/bin/preprocess script on gcc.gnu.org.

I don't see this in CVS yet.

> Joseph, do you want to do something similar for /onlinedocs?  The URLs
> generated by makeinfo 4.7 are simply gross (and currently external links
> and bookmarks into /onlinedocs are broken).

If makeinfo isn't going to change back then I think we simply need to 
update all links pointing from our pages into the onlinedocs for mainline.  
The stability of URLs for mainline docs is never guaranteed, as manual 
sections may be renamed freely.  The URLs for past releases are meant to 
be stable, so we should remember not to regenerate the onlinedocs for 
3.4.1 and before in future.  (Regeneration of past docs being something 
that could be done in principle to take advantage of makeinfo 
improvements, if it didn't change incompatibly, or of improvmements in 
update_web_docs; for the latter I regenerated the documentation from 3.1 
onwards some time back when adding generation of source and HTML tarballs 
to the script.)

-- 
Joseph S. Myers               http://www.srcf.ucam.org/~jsm28/gcc/
    jsm@polyomino.org.uk (personal mail)
    jsm28@gcc.gnu.org (Bugzilla assignments and CCs)

RE: Broken link and defunct search

[Gerald Pfeifer]

On Fri, 27 Aug 2004, Joseph S. Myers wrote:
>>   sed -e 's/_002d/-/g' -e 's/_002a/*/g'
>>
>> to a strategic place to the wwwdocs/bin/preprocess script on gcc.gnu.org.
> I don't see this in CVS yet.

I just added it locally, because it is so ugly.  If I haven't heard
back from the makeinfo folks by next week, I'll commit it.

>> Joseph, do you want to do something similar for /onlinedocs?  The URLs
>> generated by makeinfo 4.7 are simply gross (and currently external links
>> and bookmarks into /onlinedocs are broken).
> If makeinfo isn't going to change back then I think we simply need to
> update all links pointing from our pages into the onlinedocs for mainline.
> The stability of URLs for mainline docs is never guaranteed, as manual
> sections may be renamed freely.

True, but feel, and strongly so, that the new schema is a design failure 
and needlessly uglifies user-visible URLs, so I really would like us to
consider restoring the old scheme even without upstream support.

Gerald

PATCH for RE: Broken link and defunct search

[Gerald Pfeifer]

On Sat, 28 Aug 2004, Gerald Pfeifer wrote:
> On Fri, 27 Aug 2004, Joseph S. Myers wrote:
>>>   sed -e 's/_002d/-/g' -e 's/_002a/*/g'
>>> 
>>> to a strategic place to the wwwdocs/bin/preprocess script on gcc.gnu.org.
>> I don't see this in CVS yet.
> I just added it locally, because it is so ugly.  If I haven't heard
> back from the makeinfo folks by next week, I'll commit it.

Done as follows.

Installed and retested on gcc.gnu.org.

Gerald

Index: preprocess
===================================================================
RCS file: /cvs/gcc/wwwdocs/bin/preprocess,v
retrieving revision 1.38
diff -u -3 -p -r1.38 preprocess
--- preprocess	28 Aug 2003 13:05:38 -0000	1.38
+++ preprocess	5 Sep 2004 21:49:20 -0000
@@ -144,7 +144,10 @@ process_file()
              cat $STYLE > $TMPDIR/input
              printf '<set-var MHTML::INCLUDE-PREFIX="%s">\n' `pwd` >> $TMPDIR/input
              cat $f >> $TMPDIR/input
-            ${MHC} $TMPDIR/input > $TMPDIR/output
+            # Use sed to work around makeinfo 4.7 brokenness.
+            ${MHC} $TMPDIR/input \
+                | sed -e 's/_002d/-/g' -e 's/_002a/*/g' \
+                > $TMPDIR/output

              # Copy the page only if it's new or there has been a change, and,
              # first of all, if there was no problem when running MetaHTML.