The "Clone Don Porter" Project

The "Clone Don Porter" Project aims to fund full-time resources to perform work under the guidance of a Tcl Core Team member.

Attributes edit

start time: 2018-05-06
hours active: 248
currently funded: 0€
number of individual financial supporters: 0
number of organizational supporters: 0

Description edit

Under this project programmers are assigned to work full-time on improving Tcl. The goal of this project is to accomplish tasks that facilitate dgp's continued efforts as a member of the Tcl Core Team. The project relies on financial support from Tcl users, and as long as that support continues. The work queue will be curated directly or indirectly by dgp. Coordination and communication will occur primarily in the Tcl Chatroom.

Funding will be solicited from individuals and other entities with an interest in Tcl.

The first round of work for project was completed on 2018-06-24, with funding efforts still in progress.

Recommendations edit

APN I'm all for this project but sadly, can only offer moral, not financial, support! Good luck with getting the backing.

looks like a really great valgrind improvement, and if we can maintain at this level, that would be a great standard to sustain.: dgp, Tcl Chatroom, 2018-06-26

---

If you see value in this project, please add your voice here to encourage others to support the project.

Active edit

Completed edit

concatenating binary array with empty string yields the empty string: Reported and fixed.
Let TclNRTailcallEval handle namespace problems.: This is a fix for issue 3592747.
4b77647298 to sucessfully run make valgrind with no "still reachable reports": The platform for the successful run was Linux 4.16.13. If Valgrind still reports issues on your platform, please submit a report with the details.
Add tools/valgrind_suppress: Valgrind now issues no "still reachable" reports for cmdAH.test. The suppressed reports concern getpwnam_r behaviour.
Plug leak in TclSetEnv: Valgrind now issues no "still reachable" reports for env.test.
Remove dependencies between tests in env.test: This should make memory leaks easier to spot.
Fix errors in async.test under valgrind: Valgrind slows things down enough that previous assumptions about time to completion are rendered false.
Reference counting for ItclMemberCode: Similar changes were already applied for ItclObject. The larger goal is to address performance issues mentioned issue 271.
Added a gdb-test target to SampleExtension: For more convenient debugging.
Fix valgrind target in SampleExtension: TCLLIBPATH` was not being set so there was a risk of picking up a previously-installed shared object.
TIP 498, Fix memory leak due to reference counting error
TIP 498, Move refCount management of lookupNsPtr out of "resolve:/goto resolve" block
TIP 498, Fix reference counting error in TclRenameCommand
TIP 498, Fix memory leak in Tcl_DeleteNamespace
Deduplicate code in string comparison functions: Also avoid generating the string representation of pure lists and dictionaries when comparing with the empty string.
Fix memory leak in epoll: Cleanup the eventfd trigger file info created during initialization.

Waiting edit

trace: Improve error handling and resolve other known issues.
array: Redesign machinery to support array for and array searches

: c.f. http://core.tcl.tk/tcl/info/e87bcf819f03cea4

Tk: Modify to use interface of TIP 445, or demonstrate deficiencies in that TIP.
info variable introspection gap:
aspect-string-match: complete the work started by aspect to streamline and deduplicate string matching routines.

Proposed edit

TDBC::ODBC: Revise to support systems with 32-bit SQLWCHAR. See the androwish fork.
TDBC::sqlite: Reduce memory consumption for large result sets
TDBC::sqlite: Resolve known issues regarding duplicate column names in results
TDBC: Move looping constructs down into C.
Deduplicate code in string matching functions. In the same area see ticket Exponential complexity for bad patterns instead of linear.

Funding edit

bll 2018-06-25 Who the hell is "we", and who do these accounts belong to? kreidla put this information here and who is that? Scam? Does someone really expect that money would be transferred to an unknown account?

kreidla 2018-06-26: Komtere OÜ

, an Estonian firm founded in 2016 by Nathan Coulter and Karl Reidla (me), is running this project. The bank details are also available on the Komtere website

, but since the certificate for the site is self-signed, those doubting the authenticity of this information may want to contact us directly or give the information on this page sufficient time to be redacted by others if it is false.

Transactions may be made in USD and EUR via the following accounts:

USD

 Account Holder: Komtere OÜ
 Account Number: 8310069383
 Wire Routing Number: 026073008
 ACH Routing Number: 026073150
 Address: TransferWise, 19 W 24th Street, New York, 10010, United States

EUR

 Account Holder: Komtere OÜ
 IBAN: DE687001111060511738 97
 Bank code (SWIFT / BIC): EKTDE7GXXX
 Address: Handelsbank, Elsenheimer Str. 41, München, 80687, Germany

Comments edit

SYStems 2018-05-07: Can we add a Tcl/Tk modern IDE to that list :)

ak I believe that one is a step to far and big.

SYStems let's do it then :)

APN One performance enhancement I would like, if you are taking RFE's, is faster calls to commands implemented in C extensions. They are currently significantly slower than bytecoded commands.

---

[DEC] Could this be funded via Patreon?

Hacking on The Core