feat(input): :sparkles: introducing stdin support allowing to read and search any input #235 by dvorka · Pull Request #539 · dvorka/hstr

dvorka · 2025-05-18T19:14:41Z

This PR adds ability of HSTR to read input from pipe - just for inspiration:

git log --pretty=format:%s | hstr
find | hstr
ps aux | hstr
cat /var/log/nginx/access.log | hstr
grep phrase . | hstr
cat data.csv | hstr
netstat -tulpn | hstr
kubectl logs mypod | hstr

Tasks:

doc/USER_DOCUMENTATION.md Markdown documentation - added in feat(tiocsti): ✨ adding interactive "TIOCSTI not configured" warning #548 #549
man documentation
🐞 selected row not inserted: [stdin] selected rows from the pipe are just printed (not inserted) on new kernels #235 #539 #551
- 🐞 Zsh
- 🐞 Bash
documentation
- README.md/Tips - use 16b3805
  - TIPS.md - use 16b3805

Known issue:

[stdin] selected rows from the pipe are just printed (not inserted) on new kernels #235 #539 #551

…any input #235

Copilot

Pull Request Overview

This PR introduces stdin support for HSTR, allowing the utility to read input from a pipe and enable searching of piped data. Key changes include updating function signatures (e.g., fill_terminal_input and prioritized_history_create) to accept additional parameters such as file descriptors and history state, and modifying curses initialization to handle tty reading when input is piped.

Reviewed Changes

Copilot reviewed 16 out of 16 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
src/include/hstr_utils.h	Updated fill_terminal_input signature to include file descriptor parameter.
src/include/hstr_history.h	Added HISTORY_STATE parameter to prioritized_history_create.
src/include/hstr_curses.h	Modified hstr_curses_start to accept a tty initialization flag and return a FILE*.
src/hstr_utils.c	Adjusted fill_terminal_input implementation to use the new file descriptor parameter.
src/hstr_history.c	Updated prioritized_history_create to support an external history state.
src/hstr_curses.c	Updated curses initialization logic to conditionally use /dev/tty when reading from pipe.
src/hstr.c	Updated main loop and interactive functions to support pipe input and pass the correct fd.
pad.xml, hstr.pro, build scripts	Bumped version numbers and updated build scripts accordingly.
README.md	Added package information.

Comments suppressed due to low confidence (1)

src/hstr.c:1728

The code closes stdin after processing piped history input, which may be problematic if further standard input operations are expected; double-check that this behavior is intended.

fclose(fp);

Copilot

Pull request overview

Copilot reviewed 18 out of 18 changed files in this pull request and generated 6 comments.

Comments suppressed due to low confidence (1)

src/hstr_history.c:188

rawHistory is allocated with malloc and not zero-initialized. In the population loop, entries can be skipped (e.g., empty lines) without writing a value into rawHistory[rawOffset], leaving garbage pointers; later code assumes rawItems[i] is non-NULL (second pass in hstr_make_selection calls strstr(source[i], ...) without a NULL check), which can crash. Initialize rawHistory to NULL (e.g., calloc), and ensure skipped/ignored items are removed from the raw list and rawCount reflects only valid strings.

        HIST_ENTRY **historyList=historyState ? historyState->entries : history_list();
        char **rawHistory=malloc(sizeof(char*) * historyState->length);
        int rawOffset=historyState->length-1, rawTimestamps=0;
        char *line;
        int i;

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

…ation

…ck removed)

…is freeing memory

Copilot

Pull request overview

Copilot reviewed 21 out of 22 changed files in this pull request and generated 9 comments.

Comments suppressed due to low confidence (2)

.github/workflows/build_ubuntu.yml:14

The workflow runs on ubuntu-latest, but the install step is explicitly tailored to Ubuntu 24.04 packages. Since ubuntu-latest can move to a newer release and break package availability, consider pinning runs-on: ubuntu-24.04 (or adjusting the step to detect the distro version dynamically).
src/hstr_history.c:192
rawHistory is allocated with malloc() and entries are only written for non-empty lines and timestamps. If historyList[i]->line is empty (which is plausible with piped stdin input), the loop continues without setting rawHistory[rawOffset], leaving an uninitialized pointer that later gets exposed via prioritizedHistory->rawItems/rawCount and can crash when accessed. Consider initializing rawHistory to NULL (e.g., calloc), setting rawHistory[rawOffset]=NULL before continue, and/or tracking rawCount based on actually populated entries (similar to the timestamp compaction).

        HIST_ENTRY **historyList=historyState ? historyState->entries : history_list();
        char **rawHistory=malloc(sizeof(char*) * historyState->length);
        int rawOffset=historyState->length-1, rawTimestamps=0;
        char *line;
        int i;
        for(i=0; i<historyState->length; i++, rawOffset--) {
            if(!historyList[i]->line || !strlen(historyList[i]->line)) {
                continue;
            }

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-01-25T22:10:12Z

+    if(ttyInit) {
+        SCREEN *screen;
+
+        tty_in = fopen("/dev/tty", "r+");
+        if (!tty_in) {
+            perror("Could not open /dev/tty prior to curses initialization");
+            return NULL;
+        }
+
+        screen = newterm(NULL, stdout, tty_in);
+        if (!screen) {
+            fprintf(stderr, "Error initializing curses when HSTR is reading from pipe");
+            fclose(tty_in);
+            return NULL;
+        }
+        set_term(screen);
+    } else {


In the ttyInit path, newterm() creates a new SCREEN* but it’s stored only in a local variable and never released (e.g., via delscreen() after endwin()), which can leak ncurses resources. Consider storing the SCREEN* somewhere reachable by hstr_curses_stop() (or returning it alongside tty_in) and cleaning it up on shutdown.

Copilot · 2026-01-25T22:10:12Z

+Browse output of any command interactively:
+
+```bash
+ls -la | hh
+```


This new example uses hh, but earlier in the document the pipe example uses ls | hstr and other examples were switched to hstr. To avoid confusion, make the examples consistent (either use hstr everywhere, or consistently use hh and explicitly state it’s an alias).

Copilot · 2026-01-25T22:10:13Z

+		<Program_Release_Month>5</Program_Release_Month>
+		<Program_Release_Day>9</Program_Release_Day>
+		<Program_Release_Year>2025</Program_Release_Year>


The PAD release date (2025-05-09) doesn’t match the changelog entry for v3.2.0 (2026-01-25). If v3.2.0 is being released now, update the PAD date to match; otherwise align the changelog/versioning so downstream consumers don’t get conflicting release metadata.

Suggested change

<Program_Release_Month>5</Program_Release_Month>

<Program_Release_Day>9</Program_Release_Day>

<Program_Release_Year>2025</Program_Release_Year>

<Program_Release_Month>1</Program_Release_Month>

<Program_Release_Day>25</Program_Release_Day>

<Program_Release_Year>2026</Program_Release_Year>

Copilot · 2026-01-25T22:10:13Z

+    if(!historyState) {
+        using_history();
+        if(!history_mgmt_load_history_file()) {
+            return NULL;
+        }
+
+        historyState=history_get_history_state();
    }


prioritized_history_create() allocates a HISTORY_STATE via history_get_history_state() when the caller passes NULL, but it is never freed anymore. Since history_get_history_state() returns heap memory, this introduces a leak on each call. Consider tracking ownership (e.g., bool ownsHistoryState = (historyState == NULL) before assignment) and free(historyState) before returning when it was internally allocated.

Copilot · 2026-01-25T22:10:13Z

+
+        screen = newterm(NULL, stdout, tty_in);
+        if (!screen) {
+            fprintf(stderr, "Error initializing curses when HSTR is reading from pipe");


This error message is missing a trailing newline, which can make it harder to read when mixed with subsequent output. Consider adding \n (and possibly including strerror(errno)/perror style details for consistency with the /dev/tty open failure path).

Suggested change

fprintf(stderr, "Error initializing curses when HSTR is reading from pipe");

fprintf(stderr, "Error initializing curses when HSTR is reading from pipe\n");

Copilot · 2026-01-25T22:10:13Z

+  Minor release bringing ability to read input
+    from the pipe (instead of history) and filter to print it using HSTR;
+    and interactive missing configuration detection for TIOCSTI enforcement.



The release notes for v3.2.0 are duplicated/redundant (the unbulleted paragraph repeats the bullet point). Consider keeping a single concise entry to avoid confusion in downstream packaging/changelog parsing.

Suggested change

Minor release bringing ability to read input

from the pipe (instead of history) and filter to print it using HSTR;

and interactive missing configuration detection for TIOCSTI enforcement.

Copilot · 2026-01-25T22:10:13Z

+    if(!isatty(fileno(stdin))) {
+        // load history (or alternative content) from stdin
+        hstr->readingFromPipe=true;
+        int MAX_STDIN_ENTRIES = 2000;
+
+        historyState = malloc(sizeof(HISTORY_STATE));
+        if (!historyState) {
+            fprintf(stderr, "Failed to allocate memory for history state\n");
+            hstr_exit(EXIT_FAILURE);
+        }
+        memset(historyState, 0, sizeof(HISTORY_STATE));
+
+        historyState->entries = malloc(sizeof(HIST_ENTRY*) * MAX_STDIN_ENTRIES);
+        if (!historyState->entries) {
+            fprintf(stderr, "Failed to allocate memory for history entries\n");
+            free_history_state(historyState);
+            hstr_exit(EXIT_FAILURE);
+        }
+
+        FILE *fp = stdin;
+        if(fp == NULL) {
+            fprintf(stderr, "Error opening stdin\n");
+            free_history_state(historyState);
+            hstr_exit(EXIT_FAILURE);
+        }
+
+        char *line = NULL;
+        size_t len = 0;
+        ssize_t read;
+        historyState->length=0;
+        while ((read = getline(&line, &len, stdin)) != -1 && historyState->length < MAX_STDIN_ENTRIES) {
+            if (read > 0 && line[read-1] == '\n') {
+                line[read-1] = '\0';
+            }
+
+            historyState->entries[historyState->length] = malloc(sizeof(HIST_ENTRY));
+            if (!historyState->entries[historyState->length]) {
+                fprintf(stderr, "Failed to allocate memory for history entry\n");
+                free(line);
+                free_history_state(historyState);
+                hstr_exit(EXIT_FAILURE);
+            }
+
+            historyState->entries[historyState->length]->line = strdup(line);
+            if (!historyState->entries[historyState->length]->line) {
+                fprintf(stderr, "Failed to allocate memory for history line\n");
+                free(historyState->entries[historyState->length]);
+                free(line);
+                free_history_state(historyState);
+                hstr_exit(EXIT_FAILURE);
+            }
+
+            historyState->entries[historyState->length]->timestamp = NULL;
+            historyState->entries[historyState->length]->data = NULL;
+
+            historyState->length++;
+        }
+        free(line); // free(NULL) is safe per C standard
+
+        fclose(fp);
+    }


The new stdin/pipe parsing path introduces substantial logic (getline loop, max-entry truncation, blank-line handling, and memory management) but doesn’t appear to be covered by the existing Unity test suite. Consider adding unit tests around parsing piped input into a history list (including blank lines and >MAX_STDIN_ENTRIES) to prevent regressions like dangling pointers/uninitialized entries.

… and when they are, ptr is NULLed #235

…ious problems w/ file descriptors #235

Copilot

Pull request overview

Copilot reviewed 22 out of 22 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (1)

.github/workflows/build_ubuntu.yml:7

runs-on: ubuntu-latest is a moving target, but the install step is explicitly tailored for Ubuntu 24.04. Pinning to ubuntu-24.04 (or adjusting the step to be version-agnostic) will make CI more reproducible and avoid future breakages when ubuntu-latest changes.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-01-26T22:36:23Z

+    if(!isatty(fileno(stdin))) {
+        // load history (or alternative content) from stdin
+        hstr->readingFromPipe=true;
+        int MAX_STDIN_ENTRIES = 2000;
+


The new stdin/pipe ingestion path (reading lines with getline, building a synthetic HISTORY_STATE, and toggling readingFromPipe) introduces substantial new behavior but doesn’t appear to be covered by existing Unity tests. Consider adding unit tests that build a small synthetic HISTORY_STATE (including empty lines) and verify prioritized_history_create/selection logic behaves safely in pipe mode.

Copilot · 2026-01-26T22:36:24Z

        RadixItem *radixItem;
-        HIST_ENTRY **historyList=history_list();
+        HIST_ENTRY **historyList=historyState ? historyState->entries : history_list();
        char **rawHistory=malloc(sizeof(char*) * historyState->length);


rawHistory is allocated with malloc and not initialized. Later in the loop, some iterations continue (e.g., for empty lines) without writing into the corresponding rawHistory[rawOffset], leaving uninitialized pointers that can later be treated as non-NULL and dereferenced in selection/filtering. Initializing rawHistory to NULL (e.g., calloc/memset) or explicitly setting rawHistory[rawOffset]=NULL before any early-continue avoids crashes, especially now that stdin input can contain empty lines.

Suggested change

char **rawHistory=malloc(sizeof(char*) * historyState->length);

char **rawHistory=calloc(historyState->length, sizeof(char*));

Copilot · 2026-01-26T22:36:24Z


-void hstr_curses_start(void);
+FILE* hstr_curses_start(bool ttyInit);
 bool terminal_has_colors(void);


This header now exposes FILE* in the public API but doesn’t include <stdio.h>. Relying on transitive includes from curses headers is non-portable and can break compilation in translation units that include this header before <stdio.h>. Include <stdio.h> (or forward-declare struct _IO_FILE as FILE if you prefer).

feat(input): ✨ introducing stdin support allowing to read and search …

5247d2a

…any input #235

dvorka requested a review from Copilot May 18, 2025 19:14

dvorka self-assigned this May 18, 2025

dvorka linked an issue May 18, 2025 that may be closed by this pull request

Improve HSTR to read input from pipe #235

Open

Copilot AI reviewed May 18, 2025

View reviewed changes

Comment thread src/hstr_curses.c Outdated

dvorka added 4 commits May 18, 2025 21:25

fix: changing mode used to open /dev/tty #235

189ca60

build(gha): 👷 updating distro to build HSTR on GHA CI

b6f2a01

build(gha): 👷 fixing CI dependencies

8686983

build: 💚 changing apt packages for Qt on 24.04

d02987d

dvorka added the 🚀 feature label Jan 24, 2026

dvorka added 2 commits January 24, 2026 08:37

docs(adding doc to man):

b32ea0a

Merge branch 'dev-3.2.0' into enh-235/stdin-support

654e22a

dvorka requested a review from Copilot January 24, 2026 07:43

Copilot started reviewing on behalf of dvorka January 24, 2026 07:43 View session

Copilot AI reviewed Jan 24, 2026

View reviewed changes

Comment thread src/hstr.c

Comment thread src/hstr.c

Comment thread src/hstr.c

Comment thread src/hstr_history.c

Comment thread src/hstr.c Outdated

Comment thread src/hstr.c

dvorka added 9 commits January 25, 2026 20:15

Merge branch 'dev-3.2.0' into enh-235/stdin-support

4805fdb

refactor: gitignore

ff12073

fix(stdin): hide DEL

cbb5390

fix(stdin): skip favorites view in stdin mode

faf2916

build(gha): returning back GHA w/ correct config + improving document…

04f7847

…ation

doc(changelog): changelog polish.

26cc8db

fix(alloc): fixing malloc/free -- DANGEROUS change

6ff94ee

fix(stdin): preventing crash on closing pipe (IMPORTANT: previous che…

4658846

…ck removed)

fix(malloc): fixing freeing memory by changing responsibility of who …

d5b2e31

…is freeing memory

dvorka requested a review from Copilot January 25, 2026 22:02

Copilot started reviewing on behalf of dvorka January 25, 2026 22:02 View session

Copilot AI reviewed Jan 25, 2026

View reviewed changes

dvorka added this to the HSTR 3.2 - TIOCSTI strikes back! milestone Jan 26, 2026

dvorka added 2 commits January 26, 2026 22:18

fix(memleaks): ensuring that rows from pipe are not freed prematurely…

651bc0c

… and when they are, ptr is NULLed #235

Merge branch 'dev-3.2.0' into enh-235/stdin-support

3e21198

fix(stdin): removing undesired close of stdin causing potentially var…

f3df066

…ious problems w/ file descriptors #235

dvorka requested a review from Copilot January 26, 2026 22:23

Copilot started reviewing on behalf of dvorka January 26, 2026 22:23 View session

Copilot AI reviewed Jan 26, 2026

View reviewed changes

Merge branch 'dev-3.2.0' into enh-235/stdin-support

16b3805

	fprintf(stderr, "Error initializing curses when HSTR is reading from pipe");
	fprintf(stderr, "Error initializing curses when HSTR is reading from pipe\n");

	Minor release bringing ability to read input
	from the pipe (instead of history) and filter to print it using HSTR;
	and interactive missing configuration detection for TIOCSTI enforcement.

	char *rawHistory=malloc(sizeof(char) * historyState->length);
	char *rawHistory=calloc(historyState->length, sizeof(char));

Conversation

dvorka commented May 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Copilot AI Jan 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Jan 26, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 26, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 26, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

dvorka commented May 18, 2025 •

edited

Loading