Closed Bug 451150 Opened 16 years ago Closed 16 years ago

Build failed on solaris with sunstudio after the changset 606. (For bug 450126)

Categories

(Tamarin Graveyard :: Build Config, defect)

Product:
Tamarin Graveyard
Component:
Build Config
Platform:
Sun
Solaris
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(Not tracked)

Status:
VERIFIED FIXED

People

(Reporter: leon.sha, Assigned: leon.sha)

Details

Attachments

(3 files, 2 obsolete files)

Error log
9.62 KB, text/plain
Details
patch
5.26 KB, patch
stejohns
: review+
Details | Diff | Splinter Review
patch for string/namespace thunks
5.64 KB, patch
leon.sha
: review+
Details | Diff | Splinter Review
Attached file Error log —
Build failed on following files. buildin.cpp SparcAssembler.cpp NamespaceClass.cpp StringClass.cpp
Attached patch patch (obsolete) — — Splinter Review
Assignee: nobody → leon.sha
Status: NEW → ASSIGNED
Attachment #334410 - Flags: review?(stejohns)
Comment on attachment 334410 [details] [diff] [review] patch I really want to eliminate NATIVE_METHOD2, as Solaris is the only place it's needed. Can you explain why we can't have NATIVE_METHOD do the right thing?
Attachment #334410 - Flags: review?(stejohns) → review-
Using SunStudio on solaris, it is very hard to cast from one function pointer to another function pointer, especially one function is a const function. I'll try to find another way to do this.
Would it work to use reinterpret_cast<> everywhere? Or use a FUNC_CAST like you did in NativeFunction.h?
(In reply to comment #4) > Would it work to use reinterpret_cast<> everywhere? Or use a FUNC_CAST like you > did in NativeFunction.h? > reinterpret_cast can not cast const functions. Maybe I can cast off const first and use reinterpret_cast. Currently I can not try this, I'll find a better solution later. Also if I can do this, I prefer using FUNC_CAST since it will not modify other platforms.
Attached patch patch — — Splinter Review
Attachment #334410 - Attachment is obsolete: true
Attachment #346405 - Flags: review?(stejohns)
Attached patch test_patch (obsolete) — — Splinter Review
This is a patch which can make compiler happy. It is not a real patch, just to point out the problem.
Attachment #346405 - Flags: review?(stejohns) → review+
Oh, joy.... yeah, recent changes in the autogenerated stuff broke this, and I didn’t notice because most of our mainstream compilers don’t have this issue. hacking builtin.cpp directly is suboptimal, as it’s auto-generated... let me see if I can produce a proper fix in nativegen.py quickly.
I think this will do the trick -- please try it out and let me know.
Attachment #346410 - Attachment is obsolete: true
Attachment #347423 - Flags: review?(leon.sha)
Comment on attachment 347423 [details] [diff] [review] patch for string/namespace thunks This patch works good for sunstudio.
Attachment #347423 - Flags: review?(leon.sha) → review+
pushed to redux as changeset: 1079:9040b560dedc
Status: ASSIGNED → RESOLVED
Closed: 16 years ago
Resolution: --- → FIXED
Resolved fixed engineering / work item that has been pushed. Setting status to verified.
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: