Hello,
I was trying to install GNU Pascal on a 9000/735 running HP-UX 10.20. I got egcs-1.1.2 and gpc-19990409. The trouble starts early when I try to apply the patch (with patch -p1 < p/diffs/gcc-egcs-1.1.2.diff, I get:
Hmm... Looks like a new-style context diff to me... The text leading up to this was: -------------------------- |Changes for GCC version egcs-2.93.12 for GNU Pascal 2.1 | |Before applying these diffs, go to the directory gcc-egcs-2.93.12 |and use the command | | patch -p1 | |feeding it the following diffs as input. | |*** gcc-egcs-2.93.12/Makefile.in.orig Wed Mar 17 15:22:18 1999 |--- gcc-egcs-2.93.12/Makefile.in Wed Mar 17 15:24:24 1999 -------------------------- Patching file Makefile.in using Plan A... Hunk #1 failed at 2731. Hunk #2 succeeded at 2637 with fuzz 2 (offset -136 lines). 1 out of 2 hunks failed--saving rejects to Makefile.in.rej Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** gcc-egcs-2.93.12/convert.c.orig Wed Mar 17 15:22:18 1999 |--- gcc-egcs-2.93.12/convert.c Wed Mar 17 15:24:24 1999 -------------------------- Patching file convert.c using Plan A... Reversed (or previously applied) patch detected! Assume -R? [y] y Hunk #1 succeeded at 289 (offset -8 lines). Hunk #2 succeeded at 327 (offset -8 lines). Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** gcc-egcs-2.93.12/dbxout.c.orig Wed Mar 17 15:22:18 1999 |--- gcc-egcs-2.93.12/dbxout.c Wed Mar 17 15:24:24 1999 -------------------------- Patching file dbxout.c using Plan A... Reversed (or previously applied) patch detected! Assume -R? [y] y Hunk #1 succeeded at 986 (offset -2 lines). Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** gcc-egcs-2.93.12/expr.c.orig Wed Mar 17 15:22:19 1999 |--- gcc-egcs-2.93.12/expr.c Wed Mar 17 15:24:24 1999 -------------------------- Patching file expr.c using Plan A... Reversed (or previously applied) patch detected! Assume -R? [y] y Hunk #1 succeeded at 4172 (offset -170 lines). Hunk #2 succeeded at 4839 (offset -21 lines). Hunk #3 succeeded at 4703 (offset -170 lines). Hunk #4 succeeded at 6023 (offset -74 lines). Hunk #5 failed at 6033. Hunk #6 succeeded at 6452 (offset -150 lines). Hunk #7 succeeded at 6635 (offset -72 lines). Hunk #8 succeeded at 7011 (offset -150 lines). Hunk #9 succeeded at 7124 (offset -72 lines). Hunk #10 succeeded at 7171 (offset -150 lines). Hunk #11 succeeded at 7269 (offset -72 lines). Hunk #12 succeeded at 7682 (offset -149 lines). Hunk #13 succeeded at 8038 (offset -120 lines). Hunk #14 succeeded at 11026 (offset -171 lines). 1 out of 14 hunks failed--saving rejects to expr.c.rej Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** gcc-egcs-2.93.12/fold-const.c.orig Wed Mar 17 15:22:19 1999 |--- gcc-egcs-2.93.12/fold-const.c Wed Mar 17 15:24:24 1999 -------------------------- Patching file fold-const.c using Plan A... Reversed (or previously applied) patch detected! Assume -R? [y] y Hunk #1 succeeded at 1230 (offset -231 lines). Hunk #2 succeeded at 4892 (offset -33 lines). Hunk #3 succeeded at 4755 (offset -231 lines). Hunk #4 succeeded at 4960 (offset -33 lines). Hunk #5 succeeded at 4792 (offset -231 lines). Hunk #6 succeeded at 5443 (offset -33 lines). Hunk #7 failed at 6278. 1 out of 7 hunks failed--saving rejects to fold-const.c.rej Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** gcc-egcs-2.93.12/function.c.orig Wed Mar 17 15:22:20 1999 |--- gcc-egcs-2.93.12/function.c Wed Mar 17 15:24:24 1999 -------------------------- Patching file function.c using Plan A... Reversed (or previously applied) patch detected! Assume -R? [y] y Hunk #1 succeeded at 461 (offset -40 lines). Hunk #2 succeeded at 473 (offset -40 lines). Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** gcc-egcs-2.93.12/optabs.c.orig Wed Mar 17 15:22:20 1999 |--- gcc-egcs-2.93.12/optabs.c Wed Mar 17 15:24:24 1999 -------------------------- Patching file optabs.c using Plan A... Reversed (or previously applied) patch detected! Assume -R? [y] y Hunk #1 succeeded at 1383 (offset -8 lines). Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** gcc-egcs-2.93.12/stor-layout.c.orig Wed Mar 17 15:22:20 1999 |--- gcc-egcs-2.93.12/stor-layout.c Wed Mar 17 15:24:24 1999 -------------------------- Patching file stor-layout.c using Plan A... Reversed (or previously applied) patch detected! Assume -R? [y] y Hunk #1 succeeded at 67 (offset 2 lines). Hunk #2 succeeded at 104 (offset 2 lines). Hunk #3 succeeded at 121 (offset 2 lines). Hunk #4 succeeded at 986 (offset -48 lines). Hunk #5 succeeded at 1061 (offset 2 lines). Hunk #6 failed at 1313. 1 out of 6 hunks failed--saving rejects to stor-layout.c.rej Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** gcc-egcs-2.93.12/toplev.c.orig Wed Mar 17 15:22:20 1999 |--- gcc-egcs-2.93.12/toplev.c Wed Mar 17 15:24:24 1999 -------------------------- Patching file toplev.c using Plan A... Reversed (or previously applied) patch detected! Assume -R? [y] y Hunk #1 succeeded at 330 (offset -17 lines). Hunk #2 succeeded at 649 (offset -3 lines). Hunk #3 failed at 965. Hunk #4 succeeded at 1077 (offset -50 lines). Hunk #5 succeeded at 3172 (offset -228 lines). Hunk #6 succeeded at 3367 (offset -50 lines). Hunk #7 succeeded at 4232 (offset -388 lines). Hunk #8 failed at 4520. 2 out of 8 hunks failed--saving rejects to toplev.c.rej Hmm... The next patch looks like a new-style context diff to me... The text leading up to this was: -------------------------- |*** gcc-egcs-2.93.12/tree.c.orig Wed Mar 17 15:22:20 1999 |--- gcc-egcs-2.93.12/tree.c Wed Mar 17 15:24:24 1999 -------------------------- Patching file tree.c using Plan A... Reversed (or previously applied) patch detected! Assume -R? [y] y Hunk #1 succeeded at 4965 (offset 12 lines). done
And then of course the compilation in unsuccessful with parse (and other) errors in toplevel.c
Any idea what might be the trouble?
Thanks,
Dimitrios.
PS. I am not in the list, so please cc me
Dimitrios Bartzos wrote:
I was trying to install GNU Pascal on a 9000/735 running HP-UX 10.20. I got egcs-1.1.2 and gpc-19990409. The trouble starts early when I try
^^^^^^^^^^
to apply the patch (with patch -p1 < p/diffs/gcc-egcs-1.1.2.diff, I get:
Hmm... Looks like a new-style context diff to me... The text leading up to this was:
|Changes for GCC version egcs-2.93.12 for GNU Pascal 2.1
^^^^^^^^^^^^ egcs-1.1.2 is egcs-2.91.66. The distribution has a wrong symlink.
Please retry with gcc-egcs-1.1.1.diff which is the correct diff for egcs-1.1.2.
Hope this helps,
Peter
egcs-1.1.2 is egcs-2.91.66. The distribution has a wrong symlink.
Please retry with gcc-egcs-1.1.1.diff which is the correct diff for egcs-1.1.2.
Hope this helps,
Peter
Thank you, it did help, but I got stuck a little down the road. Now I get:
make[2]: *** No rule to make target `mbchar.o', needed by `gpc-cpp'. Stop.
I noticed mbchar.c being mentioned in the Changelog, but I am not sure what I should do about it. Is there a patch for it to that wasn't applied and thus ../mbchar.c was not created?
By the way, I have patch gcc-egcs-2.91.60.diff and not gcc-egcs-2.91.66.diff (and that worked giving no failures when applying the patch).
(I don't know how to use mkdiff, tried but it doesn't work).
Thanks again,
Dimitrios.
Dimitrios Bartzos wrote:
make[2]: *** No rule to make target `mbchar.o', needed by `gpc-cpp'. Stop.
I noticed mbchar.c being mentioned in the Changelog, but I am not sure what I should do about it. Is there a patch for it to that wasn't applied and thus ../mbchar.c was not created?
It should have been created by patching with gcc-egcs-2.91.60.diff. If this did not happen, please try to find out why ...
Anyway: If you manually create `../mbchar.c' with the contents "/* empty */", you will be able to proceed.
By the way, I have patch gcc-egcs-2.91.60.diff and not gcc-egcs-2.91.66.diff (and that worked giving no failures when applying the patch).
That was correct.
(I don't know how to use mkdiff, tried but it doesn't work).
`mkdiff' is for creating those diffs after hacking GPC itself. It is not needed for normal installation.
Peter
I managed to get the mbchar.* files from a newer gcc version and copied them into the gcc source directory and it looks like we got past that point.
Well, only to get this:
../../xgpc -B../../ -c -Wall -Wpointer-arith -Wmissing-prototypes -Wmissing-declarations -g -O2 `if [ x@with_shared@ == xyes ] ; then echo -fPIC; fi` --unit-path=/ece2/eeof/egcs-1.1.2/gcc/p/rts --automake /ece2/eeof/egcs-1.1.2/gcc/p/rts/files.pas /bin/sh: ==: A test command parameter is not valid. /ece2/eeof/egcs-1.1.2/gcc/p/rts/files.pas: In function `Ioselect': /ece2/eeof/egcs-1.1.2/gcc/p/rts/files.pas:162: internal error--unrecognizable insn: (insn 286 283 288 (set (zero_extract:SI (reg:SI 223) (const_int 32) (const_int 0)) (reg:SI 222)) -1 (insn_list 283 (insn_list 285 (nil))) (expr_list:REG_DEAD (reg:SI 222) (nil))) ../../../egcs-1.1.2/gcc/p/../toplev.c:1385: Internal compiler error in function fatal_insn make[2]: *** [files.o] Error 1 make[2]: Leaving directory `/ece2/eeof/gcc/gcc/p/rts' make[1]: *** [pascal.rts] Error 2 make[1]: Leaving directory `/ece2/eeof/gcc/gcc' make: *** [all-gcc] Error 2
In fact I get the warning:
/bin/sh: ==: A test command parameter is not valid.
for every pascal source file compiled.
Now, what? :)
Dimitrios.
-
Dimitrios Bartzos -
Peter Gerwinski